Test stack "speed of imagedata processing and paintcompression"

[Wilhelm Sanke]

On Mon, 04 Jun 2007, Richard Gaskin wrote:

> I had thought that you'd earlier discovered the difference between MC 
> and Rev to be simply that Rev's boot script changes the default 
> compression. This would imply that anyone who uses MC or any other 
> collection of stacks as their IDE could bring the same performance in 
> any other IDE by simply changing the paintCompression back to the 
> engine's default.
>
> What am I missing?
>
> --
>  Richard Gaskin
>  Fourth World Media Corporation



Concerning your statement and question, nothing.

I submitted the bug report after a longer discussion on the improvement 
list and had been asked by Mark Waddingham to present a test stack. The 
discussion also revealed that the Rev team was apparently unaware of 
these issues.
What astonished me in my tests was the fact that under specific 
circumstances the slowing down of imagedata under PNG compression was 
far greater than expected, namely 12 times compared to RLE when 
retrieving imagedate out of a custom property.

I thought it useful to share my findings with the Metacard list, too, 
and - although the basic fact about the boot script has been mentioned 
before on this list - as memory usually is short, I think it would not 
hurt to mention this fact again.

Wilhelm Sanke
<http://www.sanke.org/MetaMedia>